Lifeboat Foundation: Safeguarding Humanity
Toggle light / dark theme

Blog

Category: robotics/AI – Page 2

Large Language Models (LLMs) have rapidly become an integral part of our digital landscape, powering everything from chatbots to code generators. However, as these AI systems increasingly rely on proprietary, cloud-hosted models, concerns over user privacy and data security have escalated. How can we harness the power of AI without exposing sensitive data?

A recent study, “Entropy-Guided Attention for Private LLMs,” by Nandan Kumar Jha, a Ph.D. candidate at the NYU Center for Cybersecurity (CCS), and Brandon Reagen, Assistant Professor in the Department of Electrical and Computer Engineering and a member of CCS, introduces a novel approach to making AI more secure.

The paper was presented at the AAAI Workshop on Privacy-Preserving Artificial Intelligence (PPAI 25) in early March and is available on the arXiv preprint server.

Read more | >

Most computers run on microchips, but what if we’ve been overlooking a simpler, more elegant computational tool all this time? In fact, what if we were the computational tool?

As crazy as it sounds, a future in which humans are the ones doing the computing may be closer than we think. In an article published in IEEE Access, Yo Kobayashi from the Graduate School of Engineering Science at the University of Osaka demonstrates that living tissue can be used to process information and solve complex equations, exactly as a computer does.

This achievement is an example of the power of the computational framework known as , in which data are input into a complex “reservoir” that has the ability to encode rich patterns. A computational model then learns to convert these patterns into meaningful outputs via a neural network.

Read more | >

A joint research team has successfully developed a next-generation soft robot based on liquid. The research was published in Science Advances.

Biological cells possess the ability to deform, freely divide, fuse, and capture foreign substances. Research efforts have long been dedicated to replicating these unique capabilities in artificial systems. However, traditional solid-based robots have faced limitations in effectively mimicking the flexibility and functionality of living cells.

To overcome these challenges, the joint research team successfully developed a particle-armored liquid robot, encased in unusually dense hydrophobic (water-repelling) particles.

Read more | >

Scientists from TU Delft and EPFL have created a quadruped robot capable of running like a dog without the need for motors. This achievement, a product of combining innovative mechanics with data-driven technology, was published in Nature Machine Intelligence and could pave the way for energy-efficient robotics.

“Commercial quadruped robots are becoming more common, but their energy inefficiency limits their operating time,” explains Cosimo Della Santina, assistant professor at TU Delft. “Our goal was to address this issue by optimizing the robot’s mechanics by mimicking the efficiency of biological systems.”

Read more | >

Many of the robotic systems developed in the past decades are inspired by four-legged (i.e., quadruped) animals, such as dogs, cheetahs and horses. By replicating the agile movements of these animals, quadruped robots could move swiftly on the ground, crossing long distances on various terrains and rapidly completing missions.

Yet realistically and robustly replicating the fluid motions observed in animals using can be very challenging. While some existing four-legged robots were found to be very agile and responsive to changes in their environment, these systems typically integrate advanced actuators and computational components that consume a lot of energy.

Researchers at EPFL’s CREATE Lab and Delft University of Technology (TU Delft) recently developed a new four-legged robot called PAWS (Passive Automata With Synergies), which could reproduce the fluid and adaptive movements of animals using fewer actuators. This robot, introduced in a paper in Nature Machine Intelligence, leverages so-called motor synergies, which are coordinated patterns of muscle activation that allow animals to perform agile motions consuming less energy.

Read more | >

Imagine a robot that can walk, without electronics, and only with the addition of a cartridge of compressed gas, right off the 3D-printer. It can also be printed in one go, from one material.

That is exactly what roboticists have achieved in robots developed by the Bioinspired Robotics Laboratory at the University of California San Diego. They describe their work in an advanced online publication in the journal Advanced Intelligent Systems.

To achieve this feat, researchers aimed to use the simplest technology available: a desktop 3D-printer and an off-the-shelf printing material. This design approach is not only robust, it is also cheap—each robot costs about $20 to manufacture.

Read more | >